Newsvendor Bounds and Heuristics for Optimal Policy of Serial Supply Chains with and without Expedited Shippings∗

We consider N-stage serial production/distribution systems with stationary demand at the most downstream stage. First, we study classical Clark-Scarf model with both average cost and discounted cost criteria. The optimal echelon base-stock levels are obtained in terms of only probabilistic distributions of leadtime demand. This analysis yields a novel approach for developing bounds and simple heuristics for optimal echelon base-stock polices. In addition to deriving known bounds, we develop several new upper bounds for both average cost and discounted cost models. Second and more important, we extend this idea to a more general model with two transportation modes between stages: the regular and expedited shippings (Lawson and Porteus (2000)). The optimal inventory policy for this system is known to be echelon base-stock policy, which can be computed through minimizing 2N nested convex functions recursively. We again develop simple newsvendor type of lower and upper bounds for the optimal control parameters, as well as a simple near optimal heuristic. Extensive numerical results show that the heuristics for both models perform well. The bounds and heuristic enhance the accessibility and implementability of the optimal policies in supply chains with single and dual transportation modes.